Biomolecular condensates — membrane-less compartments that form inside cells through liquid-liquid phase separation — were understood as simple liquid droplets. Proteins and nucleic acids concentrate into a dense phase, separated from the surrounding cytoplasm by surface tension rather than a membrane. The interior was assumed to be disordered: a crowded solution where molecules diffuse freely. The condensate's function came from concentration, not structure.
Researchers at Scripps Research used cryo-electron tomography to look inside condensates formed by PopZ, a bacterial protein. The interior is not a simple liquid. It contains intricate scaffolds of thread-like protein filaments that self-assemble in a step-by-step process. The condensate has architecture. Published in Nature Structural and Molecular Biology, the finding extends to human cells, where filament-based condensates perform two critical functions: clearing toxic proteins (failure leads to ALS-like accumulation) and controlling cell growth (failure contributes to cancer).
The structural insight is about the relationship between boundary and interior. Membrane-bound organelles have both a defined boundary (the membrane) and internal architecture (cristae in mitochondria, cisternae in the endoplasmic reticulum). Condensates were thought to have a defined boundary (the phase interface) but no internal architecture — that was what made them different from traditional organelles. The discovery of internal scaffolding means condensates are more like traditional organelles than the phase-separation framework suggested. The boundary mechanism changed; the organizational principle did not.
This matters for therapeutics because targeting a liquid droplet is difficult — you can dissolve it or not, but you cannot modify its interior selectively. Targeting a scaffold is feasible — you can disrupt specific filament-filament interactions, modify assembly kinetics, or stabilize particular configurations. The condensate was previously a single therapeutic target. With internal architecture, it becomes a set of targets. The resolution of intervention matches the resolution of structure.
The deeper lesson is about the assumptions embedded in naming. “Liquid-liquid phase separation” implies liquid behavior throughout. The name predicted the interior would be simple, and researchers largely stopped looking. When they finally looked, using techniques capable of resolving individual protein complexes inside the condensate, they found the simple interior was complex all along.